Effect of processing parameters on the formation process of nano-sized ZrB 2 powders by the high energy ball milling

Introduction

Zirconium diboride (ZrB₂) based ceramic is well suited to extreme high-temperature environments due to its excellent properties, such as high melting point, high strength, excellent hardness, high thermal and electrical conductivity, good chemically stability, and good thermal shock resistance [1-3]. These distinctive features of ZrB₂ make it a good candidate for use in various applications, such as hypersonic flight vehicles, atmospheric re-entry vehicles, and other applications subjected to temperatures in excess of 2000°C [4,5].

However, the properties of the ZrB₂ powders (such as morphology, size, and purity, etc.) directly influence the applications of ZrB₂ ceramic material. The nanopowders help improve the microstructure of the sintered ceramic sample, and can thus enhance its properties, such as the densification, mechanical strength, thermal, and electrical conductivities, which are application sensitive [1,6-8]. Hence, selecting a suitable method for preparing nano-sized ZrB₂ powders is of great importance. Currently, nano-sized ZrB₂ materials are usually fabricated using several preparation methods, such as the carbothermic reduction of ZrO₂ and B₂O₃ [9], sol-gel reactions [10,11], self-propagation high-temperature synthesis (SHS) [12,13], and the direct high-temperature sintering of metallic Zr, its oxide, or hydrides with elemental boron [14]. However, it is always necessary to have a high temperature (>1500°C) and a relatively complex and long production period in those methods, and the synthesised powders usually have a relatively large particle size. Therefore, a reduced energy and simplified process for synthesising ZrB₂ nanopowders is needed.

In recent years, nano-sized boride powders have been synthesised by a high energy ball milling method that includes metallic Mg powder among the precursor powders [15,16]. This method relies on a self-sustaining reaction generated by the highly exothermic magnesiothermic reduction of oxide precursors induced by high energy ball milling. Thus, the preparation process can be carried out at room temperature without any extra heating steps. Luo et al. [16] synthesised nanoscale ZrB₂-TiB₂ composite powders via planetary ball milling for 120 h. Setoudeh and Welham [17] synthesised ZrB₂ by conducting 15 h of ball milling using ZrO₂, B₂O₃, and Mg powders as starting materials. In contrast to the conventional SHS procedure [12], the ball milling-assisted synthesis method provides numerous advantages. For example, the processes of reactant mixing, materials synthesis, and particle size reduction can be conducted in a single step, which greatly simplifies the synthesis process and reduces energy requirements. Thus, the ball milling-assisted synthesis process combines relatively low energy, low cost, and simplicity with mild reaction conditions and high yield. However, the synthesis reaction rate and the resulting particle size can be strongly influenced by the ball milling parameters, such as the milling speed, milling time, and milling ball medium to starting powder mass ratio. There is still a lack of systematic study about how the milling ball process affect the milling ball process affect the self-sustaining reaction of ZrB₂ nanoparticles.

The present study capitalises on the advantages of the high energy ball milling-assisted synthesis process to prepare nanoscale ZrB₂ powders by a self-sustaining reaction using ZrO₂, B₂O₃, and Mg powders as starting materials. The influence of the molar ratios of ZrO₂, and B₂O₃, the precursor powder to milling ball medium weight ratio, the characteristics of the milling medium, and reaction time, on the phase structure, morphology, and chemical component of the powders is investigated systematically to determine how ball milling process parameters affect the chemical reaction course and resulting product quality.

Materials and methods

Results and discussion

Figure 1(a,b) present XRD patterns of the products obtained after 8 h of high energy ball milling with steel balls or ZrO₂ balls as the milling media, respectively, and with various ZrO₂:B₂O₃ molar ratios. As shown in Figure 1(a), the as-milled products obtained using steel balls as the milling medium included ZrB₂ (ICCD PDF 34-0423) and MgO (ICCD PDF 45-0946) as the major phases when the Zr:B molar ratio was 1:6, indicating the occurrence of a self-sustaining synthesis reaction. In addition, a minor phase of magnesium borate (Mg₃B₂O₆, ICCD PDF 38-1475) was contained in the products. However, any of the other Zr:B molar ratios employed failed to meet the basic reaction conditions for the synthesis of ZrB₂ using a steel ball milling medium, and only the starting materials, such as monoclinic ZrO₂ (ICCD PDF 13-0307) and Mg (ICCD PDF 35-0821), were observed in the products. The absence of peaks for B₂O₃ is not unexpected because the starting B₂O₃ powder provided no discernible XRD peaks. As shown in Figure 1(b), the as-milled products obtained using the less dense and smaller diameter ZrO₂ balls as the milling medium included a small concentration of ZrB₂ and MgO even if the Zr:B molar ration was as low as 1:2. Moreover, some concentrations of monoclinic ZrO₂ residual to the reaction process and tetragonal ZrO₂ (ICCD PDF 02-0733) derived from the ZrO₂ balls are present in the products. As the relative B₂O₃ content increased, the diffraction peak intensities of ZrB₂ and MgO increased significantly, and the diffraction peaks associated with monoclinic and tetragonal ZrO₂ nearly disappeared. As the Zr:Br molar ratio increased to 1:6, the diffraction peaks associated with ZrB₂ and MgO gradually weakened and broadened, and the diffraction peak intensity of tetragonal ZrO₂ increased significantly. These results illustrate that the molar ratio of ZrO₂ to B₂O₃ in the starting material is a critical processing parameter in the synthesis of ZrB₂ powders using the high energy ball milling method.

OPEN IN VIEWER

Figure 1

XRD patterns of ZrB₂ products prepared with different Zr:B molar ratios by milling for 8 h (a) using a steel ball medium when the ball-to-powder weight ratio is 40:1 and (b) using a ZrO₂ ball medium when the ball-to-powder weight ratio is 30:1.

The magnesiothermic reduction process for ZrB₂ involves a mechanism including three sub-reactions [20,21] by which the B and Zr components are separately reduced from their corresponding oxides by Mg, and combined to form ZrB₂. These sub-reactions and their corresponding changes in Gibbs free energy (ΔG) and enthalpy (ΔH) at 298 K, as well as the adiabatic temperature (T_ad) are given as follows. (2) (3) (4)

Although all of these reactions are exothermal, the overall process can be self-sustaining only if T_ad is greater than 1800K [15]. Thus, Reactions (2) and (4) are self-sustaining reactions, while Reaction (3) is not, and extra heat is necessary to induce this reaction. Under circumstances where ZrO₂, B₂O₃, and Mg are simultaneously mixed together, the overall reaction process relies on the heat generated by the highly exothermic Reactions (2) and (4) to stimulate the reduction of ZrO₂ to Zr in Reaction (2), and thus make the overall reaction process self-sustaining. Therefore, the relative B₂O₃ content is sufficient for Reaction (2) to release sufficient heat to drive Reaction (3) only for Zr:B molar ratios of 1:6 in the case of a steel ball milling medium (Figure 1(a)), and 1:4 in the case of a ZrO₂ ball milling medium (Figure 1(b)), and thus provide a self-sustaining magnesiothermic reduction. However, if the B₂O₃ content is too high, the volume of starting material will be dramatically increased because the density of B₂O₃ (2.46 g cm^–3) is much lower than ZrO₂ (5.89 g cm^–3), and thus decrease the milling energy even if the milling ball to starting powder weight ratio is kept constant, which is not beneficial for the formation of ZrB₂.

To further investigate the impact of synthesis parameters on the above synthetic reaction, the XRD patterns of as-prepared powders obtained after different milling times are shown in Figure 2(a,b) with steel balls and ZrO₂ balls as the milling media, respectively. These results are discussed separately as follows.

OPEN IN VIEWER

Figure 2

Effects of milling time on the phase composition of the products prepared using a steel ball medium with Zr:B = 1:6 (a) and a ZrO₂ ball medium with Zr:B = 1:4 (b), and on the crystallite sizes (c) and ZrB₂ mass fractions (d) obtained from the XRD data.

For the as-prepared powders obtained using steel balls as the milling medium, it is clear that the major phases are the starting monoclinic ZrO₂ and metallic Mg after milling for 6 h. The absence of diffraction peaks for MgO and ZrB₂ indicates that the shorter milling time fails to achieve the conditions required for Reactions (2)–(4). Extending the milling time to 8 h results in a large amount of synthesised MgO and ZrB₂ in the prepared samples, and a small Mg₃B₂O₆ content can be observed. This demonstrates that a mechanically induced self-propagating reaction is initiated during the high energy ball milling process. With further increase in the milling time, all of the peaks for the ZrB₂ and MgO phases gradually broaden and decrease in intensity, indicating a reduction in the average crystallite size, a buildup of defects, and the formation of internal strains [22]. This development is illustrated in Figure 2(c). In addition, the mass fractions of ZrB₂ in these samples were calculated as a function of milling time, and the results are given in Figure 2(d). The figure demonstrates that 95.8% of the ZrO₂ starting material by mass has been converted into ZrB₂ after milling for 8 h. Increasing the milling time to 12 h results in an apparently complete reaction with only peaks observed for MgO, ZrB₂, and Mg₃B₂O₆.

For the samples prepared using a ZrO₂ ball milling medium, small MgO and ZrB₂ contents were generated after just 4 h of milling, although large Mg and ZrO₂ contents remain. Here, the mass fraction of ZrB₂ is only 25.2%. However, increasing milling time is favourable to the formation of ZrB₂, and the mass fraction of ZrB₂ increases to 93.2% for a milling time of 8 h. Ultimately, the mass fraction reaches 100% after milling for 12 h. Similar to the results obtained for the samples prepared using a steel ball medium, these samples also present only trace amounts of the Mg₃B₂O₆ phase, although the peak intensities associated with the Mg₃B₂O₆ and MgO phases are relatively low. In addition, the crystallite size of ZrB₂ also presents a decreasing trend for samples prepared using a ZrO₂ ball milling medium. However, the crystallite size is greater in this case than for the samples prepared using the steel ball medium.

A comparative analysis of these results indicates that the formation of ZrB₂ is affected by the milling medium. Using relatively low density and small diameter ZrO₂ balls is beneficial for decreasing the required synthesis time, but the ZrB₂ yield is relatively low. Conversely, the use of a steel ball milling medium is advantageous for the synthesis of ZrB₂ due to the higher milling energy resulting from the greater mass of the steel balls, despite the longer synthesis period required. Thus, a higher mass fraction and smaller crystallite size can be obtained from high energy ball milling with steel balls for 8 h.

The synthesis pathway of ZrB₂ during the milling process was analysed by investigating the effect of milling time on the median particle diameter (D50) of the ZrB₂ powders prepared using a steel ball medium with Zr:B = 1:6 and a ZrO₂ ball medium with Zr:B = 1:4 (Zr:B = 1:6 and Zr:B = 1:4 are the standard ratios employed hereafter for the two milling media), and the results are presented in Figure 3. As the figure shows, the median particle diameter could be influenced greatly by the milling medium and milling time. With the steel ball medium, the median particle diameter of the samples increased slightly with increasing milling time at first, and attained its maximum D50 particle diameter (55.5 μm) after only 6 h of ball-milling treatment. However, as the milling time increased, the D50 particle diameter decreased gradually, and attained a value of only 1.9 μm after milling for 12 h. In contrast, employing the smaller and less dense ZrO₂ ball milling medium produced a more highly variable trend in the particle diameters of the samples. As shown in Figure 3, the median diameter is reduced to 2 μm after only 6 h. However, continued ball-milling processing only slightly decreases the median particle diameter due to the secondary agglomeration of the particles.

OPEN IN VIEWER

Figure 3

Effect of milling time on the median particle diameter of the milled products prepared using a steel ball medium with Zr:B = 1:6 and a ZrO₂ ball medium with Zr:B = 1:4.

These results can be combined with the SEM micrographs given in Figure 4(a,b) for products obtained after ball milling for 6 h using a steel ball medium and a ZrO₂ ball medium, respectively, to evaluate the effect of the milling medium on the synthesis pathway of ZrB₂ during the milling process. First, we note that the influence of the milling medium on the median diameter of the samples may be attributed to the different milling energies and efficiencies obtained using milling balls with different diameters and densities. The greater density and diameter of the steel balls provide greater momentum and impact force during the high energy ball milling process, and this leads to a significant degree of cold welding between the oxide powders and Mg. As a result, the Mg, B₂O₃, and ZrO₂ starting materials can form large particles with diameters of several tens of micrometers, as shown in Figure 4(a). Under this condition, the milling efficiency is greatly reduced and Reaction (2) cannot be initiated due to insufficient reaction energy. Thus, neither ZrB₂ nor the intermediate products are generated when the milling time is less than 8 h. Conversely, the lower density and smaller size of ZrO₂ balls provide a relatively low momentum and impact force that are not sufficient to cause cold welding between the reactant particles. In addition, the greater number of ZrO₂ balls employed provides a much greater crushing effect and milling efficiency, and the sizes of the reactant particles are continually reduced over a relatively short milling time, as shown in Figure 4(b). Thus, the large number of newly formed surfaces increases the overall activity of the reaction system, which facilitates the gradual generation of Reactions (2)–(4), even for milling times of only 4 h.

OPEN IN VIEWER

Figure 4

SEM micrographs indicative of the surface morphologies of products obtained by ball milling for 6 h using a steel ball medium (a) and a ZrO₂ ball medium (b).

Figure 5(a,b) present representative TEM images of ZrB₂ powders ball-milled for 12 h using steel ball and ZrO₂ ball media, respectively. Clearly, the high energy ball milling processes have reduced the particle diameters to a nanometer scale of approximately 200–400 nm. Furthermore, the higher magnification TEM images given in the insets of the figures indicate that the ZrB₂ powders are actually porous agglomerates consisting of many individual small particles with an average diameter of ∼50 nm. According to EDS results given in Figure 5(c), the oxygen contents of the ZrB₂ powders ball-milled for 12 h using steel ball and ZrO₂ ball media were 2.69 and 1.23 wt-%, respectively. The oxygen content data of these samples accords with the typically reports [23-26], which are normally in the range of 1–3 wt-%. The low oxygen contents also verify that the synthesis reactions were conducted to completion. In addition, the element of Mg (≤0.67 wt-%) could be observed from the EDS data because trace amounts of residual MgO in the acid-etched samples. And the element of Cu was derived from the copper sample stage of TEM.

OPEN IN VIEWER

Figure 5

TEM images for the acid-etched ZrB₂ samples obtained by ball milling for 12 h using a steel ball medium (a) and a ZrO₂ ball medium (b). In addition, EDS data is presented in (c) for the samples given in (a) and (b).

EDS is not strictly a surface analysis technique owing to its large interaction volume. To further study the oxidation behaviour during the high energy ball milling process, the prepared ZrB₂ powders were investigated by XPS. The XPS Zr 3d and B 1s spectra of sample surfaces obtained by ball milling for 12 h using a steel ball medium and a ZrO₂ ball medium are presented in Figure 6. The results demonstrate the presence of two Zr 3d_5/2-3d_3/2 doublet peaks and two B 1s singlet peaks for each sample, which represent the two different chemical states for the Zr and B atoms. Here, both samples present Zr 3d_5/2, Zr 3d_3/2, and B 1s peaks located at binding energies of ∼178.5, 180.8, and 187.1 eV, respectively, which are indicative of ZrB₂ [27,28]. The other set includes Zr 3d_5/2, Zr 3d_3/2, and B 1s peaks at higher binding energies of ∼182.2, 184.6, and 192 eV, respectively, which are indicative of ZrO₂ and B₂O₃ [27,28]. The oxides can also be confirmed by the obvious O 1s peaks at 531.6 eV, as shown in Figure 6(b). Therefore, the XPS results clearly indicate the presence of ZrO₂ and B₂O₃ on the surfaces of both samples. Furthermore, it can be also observed that the ZrO₂-related peaks encompass greater areas than the ZrB₂-related peaks. This observation, when taken together with the very low oxygen content data in Figure 5(c) obtained from a relatively large interaction volume by EDS and the fact that XPS is a surface analysis technique, indicates that most of ZrO₂ and B₂O₃ cover the surfaces of ZrB₂ particles [28].

OPEN IN VIEWER

Figure 6

XPS spectra of the acid-etched ZrB₂ samples after milling for 12 h.

To confirm this hypothesis, the samples were further examined by the high-resolution TEM. As can be seen in Figure 7, the direct high-resolution TEM images indicate the presence of an amorphous nano-film at the edges of the agglomerates (marked with red lines). The surface oxidation of ZrB₂ particles can be mainly attributed to the following two conditions [28]. (1) A lower chemical affinity between the oxide crystals and ZrB₂ crystals than between ZrB₂ crystals themselves promotes cold-welding during high-energy ball-milling via contacts of the type ZrB₂–ZrB₂. (2) The ball-milled particles are subjected to spontaneous surface passivation when exposed to ambient conditions.

OPEN IN VIEWER

Figure 7

High-resolution TEM bright-field images of the edges of the acid-etched ZrB₂ powder particles obtained by milling for 12 h and using a steel ball medium (a) and a ZrO₂ ball medium (b).

Conclusions

Nano-sized ZrB₂ powders have been synthesised from a starting mixture of ZrO₂, B₂O₃, and Mg powders via the high energy ball milling. The suitable ratio of ZrO₂ to B₂O₃ is critical to ensure that the overall synthesis reaction was self-sustaining. An excess of B₂O₃ can provide the release of sufficient heat through a self-sustaining magnesiothermic reduction. However, when the ratio achieves its critical ratio, increasing B₂O₃ dramatically adds the volume of reactants, and weakens the milling efficiency, which is detrimental for the synthesis of ZrB₂. Moreover, the milling efficiency can also be affected by the milling media and milling time, which will ultimately influence the synthetic process of ZrB₂. A relatively low density and small size ZrO₂ ball medium is beneficial for decreasing the milling time required for ZrB₂ synthesis, and a small amount of ZrB₂ (25.2 wt-%) can be synthesised even after milling for only 4 h. However, as such, the yield is relatively low. Conversely, a greater density and larger diameter steel ball medium provides greater momentum and impact force during the high energy ball milling process, which causes the Mg, B₂O₃, and ZrO₂ raw materials to become cold welded together at the beginning of the process. The resulting agglomerated microparticles significantly reduce the milling efficiency and defer the inception of a self-sustaining reaction. Nevertheless, the higher milling energy of this preparation condition ultimately provides a greater yield and smaller crystallite size for the final product. The average diameter of the obtained ZrB₂ powders is about 200–400 nm, which are an agglomeration composed of many individual small particles with an average diameter of ∼50 nm. And the surface of the resulting particles would be oxidised lightly.